装备电路组件众多且结构复杂,多组件复合故障程度难以准确评估,导致装备运维方案无法有效规划.针对该问题,以剩余寿命作为故障程度的分级标准,提出了一种基于剩余寿命预测的装备电路故障程度分级方法.针对装备电路在实际服役过程中获取大量故障数据样本困难的特点,提出了一种基于数模融合的寿命预测方法,利用Multisim建立故障仿真模型来扩充故障数据集,考虑到故障数据的时序性与非线性,采用双向长短期记忆网络(bidirectional long short-term memory network,BiLSTM)构建剩余寿命预测模型;构建基于剩余寿命灰色关联分析的故障程度分级模型,通过量化装备电路不同剩余寿命下多级故障程度的阈值,识别装备电路的故障程度等级;以某装备电路为例验证了所提方法的有效性.
Classification Method for Circuit Fault Severity of Certain Equipment Based on Residual Life Prediction
Equipment circuits consist of numerous components with complex structures,making it dif-ficult to accurately assess the severity of multi-component composite faults,which impedes effective plan-ning of equipment operation and maintenance schemes.To address this issue,a grading method for equip-ment circuit fault severity based on residual life prediction is proposed with the residual life as the criterion for fault severity classification.Firstly,to overcome the scarcity of fault data samples in real-service sce-narios for certain equipment types,a life prediction approach based on digital-analog fusion is introduced.A fault simulation model is established with Multisim to augment the fault dataset.Given the temporal and nonlinear nature of fault data,a bidirectional long short-term memory network(BiLSTM)is employed to construct a residual life prediction model.Secondly,a fault severity classification model based on resid-ual life grey correlation analysis is developed to quantify the multi-level fault severity under different resid-ual life scenarios and identify equipment fault severity levels.Finally,the efficacy of the proposed method is validated through a case study involving a specific type of equipment.
NETL
NSTL
万方数据
